Alienware 17 r2 980M/4980HQ

Good evening guys I've just received my refurbished Alienware notebook, initially it was completely dead the Laptop wouldn't even post to CPU damage apparently and Dell sent and engineer to my offices to repair it which worked absolutely fine. However since then I've been using it quite a bit and noticed how hot it's getting.

So after some research people were saying 240W PSU was absolutely needed to ensure it can deliver enough power to the laptop. So I'm using the 240W PSU now however my CPU is currently sitting at like 70 under very little load if I leave it completely idle it goes to around 65 and underload about 80 degrees however this is only because of the A00 bios that I reverted the machine to as A05 was getting the CPU close to around 94 degrees.

I was hoping somebody out there has worked out the best combination of cooling this and maintaining nice temperatures perhaps a third party tool to set a fan profile as Alienware don't supply one (Shocked)

I have to say I haven't been impressed if I had bought this for 2.2k from Dell it would be going straight back but because I only paid 1.3k it seems like I should give it a go. Also the sound isn't the best from this keeps stuttering and juggering. I'm hoping there's somebody out there with a laptop like mine and has got it working nicely for them as currently this laptop runs far too hot for my liking. I'm going to run 3DMARK firestrike also and will post the results back with temperatures and scores to see if you think they're normal.

Thanks for reading my thread and input would be highly appreciated

http://www.3dmark.com/3dm/8287159

Results GPU didn't go above 68 Degrees which is a good sign

However the CPU as a package went as high as 95 Degrees. Which would suggest it either started to throttle (These temperatures are far too high) also on the right hand heatsink there's an awful rattling sound would be bug the hell out of me what should I do guys? It would appear that the Laptop can't cope?

 

Your 3DMark score is lower than it should be I have the 4720HQ/980m and I get around 8350-8450; around 8700-8800 with a slight overclock to GPU. I'd suggest you to use HWiNFO 64 to manually control the fans and you should check if your CPU throttles. My CPU throttles down to 2.6 GHz if I don't use 3rd party applications such as Throttlestop.

 

I'm not sure how you manually control the fans through HWiNFO64 and have zero experience with Throttlestop could you help me out mate?

I'd personally rather hear some fans going rather than it just heating up and up like it's doing now.

Any help will be highly appreciated

 

https://www.dropbox.com/s/p2ee6s9z8xd7zfr/Log.CSV?dl=0

Doesn't seem to be throttling the core clock at all.

However one of the cores reaching 96 Degrees is a bit of a concern for me

How can Dell/Alienware get away with selling these systems without fan control?

 

I tried using HWiNFO64 however disabling ECC support leaving the fans as Auto like they have been before I then experience really bad frame rate stutter in 3DMARK, Is this a return to dell jobby?

 

Firstly, are you still on Bios A00? If so, good. Consider waiting for A06 (and for others to test it) before flashing it on your system - won't be too surprised if it's another dud as usual (it's been 5 dud bios already, what's 1 more).

Personally I wouldn't use HWInfo to solve something that the bios should handle properly in the first place. If you want additional cooling, undervolting it with XTU is a good and safe option to reduce heat under load, short of a repaste (ThrottleStop works too, but I prefer XTU's interface). Refurb quality is questionable most of the time since minor things sometimes gets overlooked by OEMs, but you should consider a return anyway if it really disturbs you (it might come back with something worse, been there done that ).

 

Yeah I'm currently on A00 bios and my hottest core through 3DMARK was 96 which in my opinion they shouldn't be selling a machine with a cpu they can't cool? The CPU heatsink/fan is rattling also (maybe the engineer didn't fit it probably when he came and replaced the whole board last week as it wouldn't boot. Seems like the GPU is absolutely fine temperature and clock speed wise actually have some headroom there)

Ok I've completely got rid of HWInfo as it was concerning me anyhow. I really do appreciate your suggestions but I don't feel as if it's something I should have to do to get the system working functionaly without it overheating shouldn't alienware have this done before they sell these machine (sorry mate not having a go at you)

I'll get the Dell engineer out again to look at the rattling CPU heatsink/fan and also the Audio is really choppy like cuts out every 10 seconds on the speakers so need to get them replaced well I've wasted so much time on the machine trying to get it fixed/right I'm kinda hoping Alienware do the right thing and supply me with a brand new unit?

What would you do in my shoes? Send it back and go for a different brand of laptop with proper fan control?

 

All of my settings are at default - https://www.dropbox.com/s/plyn7m9976v9yhj/Untitled.png?dl=0

Not sure whether undervolting would change temperatures that much what'd you think?

 

No worries, I'd be equally pissed if I was in your shoes with those temps (that said, undervolting the CPU is a a quick and easy way to get it cooler in the meantime while you resolve issues with Dell - just don't mention it to non-techs, they don't get it). In any case, some other people seem to have similar problems with heat, and more others (like me) without. So it's not likely to be a systematic problem with the lineup.

It seems you have the i7 4980hq, which turbo boosts up to the 4ghz range. That might be the reason behind it. If it isn't, then there are 3 other possibilities over here:
1) the thermal paste wasn't applied well
2) the rattling noise is related to the cooling problem
3) defective hardware (not my area of expertise, I deal with software)

One way that will definitely drop the temps is lowering all the active core settings under "Multipliers" until it reaches acceptable CPU temperatures, as a temporary stopgap measure.

 

Yeah it's definitely turbo into 4Ghz which I agree it's causing the problem in my opinion however the Chip should be capable of this with good enough cooling.

How much should I undervolt by that's going to make my CPU unstable what sort of multipler should I set the 4 cores to like 3? See if the temperatures drop drastically? The board was replaced with a brand new unit so I suspect maybe it's just the cooling unit that's failing perhaps it doesn't seem to be seated properly with the rattling noises. I'll give them a call and see what they say thanks for your help dude.

 

Ok I just set the fans to maximum through HWInfo and loaded 3DMARK the frame stuttering still happening however it would appear the GPU is making the funny noise almost like coilwhine

However when I reset them all back to default fan profiles the fan on the right hand side starting clicking like it was touching something it shouldn't time to get the dell engineer out I believe get him to repaste it properly and seat everything properly.

 

There's the 17R2 benchmark thread, so it might contain information on what temps to expect. Having an AW15, I don't follow that thread much. But my i7 4710hq runs at 3.5ghz on 4 cores at a very acceptable temperature, so I believe you could clock yours down to 3.5 - 3.8ghz for better temperatures until they fix it (and I'm guessing the same thing as you right now).

Most CPUs should undervolt -50mV easily, -70mV shouldn't be too hard, whereas good ones can hit -90mV to -100mV (maybe even more, although less likely). It's easy to get into it, but incredibly hard to perfect. Expect lots of restarting while tuning it to better undervolts, but basically few degrees shaved off.

Each 1x you reduce is called a "bin", which translates to a 100mhz underclock to CPU core speeds. So in your case, 35x for 1 Active Core means 3.5ghz for single core, and 33x for 4 Active Cores equals 3.3ghz when all 4 cores are running. Just adjust to your preference here. Temperature reduction in this case can be great, but always comes at the cost of performance.

 

Can't seem to find anyone with the 980m like myself with the 4980HQ

Will contact Dell and see what they say.

 

All defaults 4 active cores should be at 3.8Ghz however when doing the XTU stress test cpu usage ranging from 97-99% understandably temperatures never really exceed 83 degrees currently.

However Core frequency is between 3.2Ghz and 3.4Ghz which clearly shows throttling.

What should I say to Dell? can somebody come out to stop the throttling please? Surely their guys should do all of this when testing a machine?

Getting close to just saying F it and just sell it

 

Having 95'C on stock and a rattling noise there: that's a good enough reason for them to take a look at it. Btw, if you lowered the clocks earlier, then it's normal to have a lower core frequency. P.S: however if you think there's throttling going on, it can be determined by enabling these options in XTU before running a CPU stress test:

 

Cheers dude, yeah as soon as I put those on, Power limit throttling went straight to 100%

Contact Dell and hope they can sort it?

Does my graphics score seem about right on the 3dmark score I linked also?

GPU didn't go above 70 degrees so that should be fine it's just this bloody CPU kinda wish they'd give me £400 back and put a 4710 in there now

 

Of course doing this test it's not allowing it to go to 4Ghz due to power limit throttling which in turn would lead to Thermal throttling I guess.

Weird that in 3D Mark it seems to stay at 3.8Ghz when doing the CPU testing however with temperatures of 96 you can see why

FFS

 

Nah, actually that's perfectly normal. You see, whether it's a i7 4710/20hq or a 4980hq, the TDP limit is still 47w (set by Intel). No matter which OEM you go, it's the same problem unless you get a socketed CPU (which means something like a Clevo or AW18).

Undervolting reduces the temps under load AND reduces the wattage used to perform tasks at a certain speed. For me, without undervolt my system would power limit throttle under load to 3.2-3.3ghz. After, it goes straight to 3.4-3.5ghz. Increasing the TDP limit itself is simply setting the "Turbo Boost Power Max" setting to something else, like 58w, however it is not guaranteed to work indefinitely on all HQ chips (mine works, but needs a restart).

The reason it isn't throttled in 3dmark is because the CPU isn't utilized 100% (it's meant to simulate an extreme game, and games don't hit 100% most of the time, if at all). As for the higher temperatures, the CPU and GPU share the same heatsink, so heat from one transfers to the other (and if both are hot, well...). It's still uncomfortably high though, regardless.

 

Indeed, I mean running 3DMark is like a 4 minute benchmark to see my CPU hit 93 Degrees is a concern if I was playing witcher 3 on high/ultra settings I can't see this laptop not BSOD, or just shutting down due to temperatures

Hopefully Dell can sort it as apart from the temperatures I love this Laptop, can't find how to use the webcam though as there's no application on here to use it

 

Well... you could always underclock the aforementioned multipliers temporarily - from your earlier remarks it seems you haven't done so, and now's a good time as any (just remember to restore to default when the technician comes along, which you can find under Profiles tab > "Default" and click Apply) Windows 8.1/10 should have the Camera app, you can search for it in the list of apps installed.

 

When I go to the Camera app it says "To get started, connect a camera" I guess that's another thing for the engineer to look at.

I think I'm going to leave it mate. see what they say. I kinda bought the 4980 so it'd run at 3.8Ghz not say 3

Which part do I undervolt if I do? both parts? I'll probably go -40MV on the ones you suggest. Won't mess around too much with it

Once again thanks for all of your replies.

 

I get your point there - why pay the premium for the 4980, if the 4710 at 3.3ghz does a pretty job too? Just a temporary problem until they fix it, so that's a temporary solution - and you wouldn't have to drop it much, only enough to be comfortable: from 3.8ghz on 4 cores, to say 3.5ghz on 4 cores. It should be significantly cooler, and still very fast.

It's actually beneficial for undervolting to do underclocking first, since a lower performance ceiling increases the likelihood of a better undervolt (more performance = needs more voltage = less undervolt). In my case, I can only undervolt to -70mV at 3.5ghz (4 cores): with 3.3ghz (on 4 cores, which is actually stock), I'm pushing -90mV.

For starters, you would want to undervolt the "Core Voltage Offset" to -50mV. Drop it lower in steps of -10mV until it bluescreens or restarts, then you know you've hit the limit there. So if it crashes at -80mV for example, dial it back to -70mV, and run some stress tests.

For added stability, you should keep it locked at high performance mode. Or instead (for battery life), leave it idle or playing some Youtube video on loop overnight, to find out what works. Undervolting is adaptive, which means it applies whether or not it is under load, so it's necessary to test both high and low frequencies (unless you remove the possibility of the latter).

As for why it's not out-of-the-box, and if it's safe, why don't manufacturers do it...
Undervolting is pretty much like a lottery, all chips are made differently, and neither Intel nor the OEM will spend the time testing this additional benefit for you. That's why it doesn't come out of the box - it's time consuming, and people will find it unfair if they have a less stellar chip on their hands. Besides, most people don't care, so it's easier to set a baseline voltage, and call it a day. Any undervolt benefit you get, is purely good for your chip. Have fun cheers mate

Edit: in addition, you can undervolt the "Cache Voltage Offset" (which is your CPU cache) and the "Processor Graphics Voltage Offset" (the Intel iGPU), although the gains won't be as obvious. It's better to undervolt them separately, however I find cache usually undervolts as well as CPU voltage, or just needs to be 5~10mV higher, YMMV.

 

Think what the plan is to call Dell and get them to confirm the temperatures are far too high.

after some 3dmark testing CPU was reaching 96

I'll get them to replace the heatsink's and I'll supply them with IC Diamond Paste so hopefully the temperatures will be sorted soon.

 

You mean it's still reaching 96'C, even after underclocking and undervolting? (the main goal being able to game in the meantime without heat issues)

 

Nah I'm not going to underclock/undervolt until it's sorted mate not even going to use it until it's sorted

 

I did some light testing so you can see expected gains (probably more in your case, since the effect is exponential in nature). Maybe you'll reconsider that decision

Stock settings:


Undervolt (-120mV)


Undervolt (-120mV) + underclock (-300mhz)